GPS/惯性组合导航系统的并行算法研究  

A Parallel Algorithm for GPS/Inertial Integrated Navigation System

作　　者：慕德俊[1] 佟明安[1] 戴冠中 

机构地区：[1]西北工业大学电子工程系,陕西西安710072

出　　处：《西北工业大学学报》2000年第2期208-211,共4页Journal of Northwestern Polytechnical University

基　　金：西北工业大学"双新计划"资助

摘　　要：基于导弹发射段 GPS/惯性组合导航系统模型 ,研究并行化滤波计算的平方根算法、U- D算法及实现这些算法的阵列结构 。The combination of GPS and inertial navigation system (INS) during the firing stage of a medium or long range missile [2] requires so much computation in each sampling interval that the present computers experience computational bottleneck. We propose square root covariance and U D decomposition integrated filter algorithms which are suitable for computation in multiphase systolic arrays. Because the state and measurement matrices are sparse matrices, the square root covariance and U D decomposition filter algorithms with sub matrices are proposed to reduce the amount of computation and systolic arrays. The U D filter algorithm is free from the square computations and can improve numerical stability. The analysis and simulation result with a parallel simulation software——PVM, shows that the speed up ratio for U D filter algorithm is 48. Therefore, the algorithms and systolic arrays proposed in this paper can meet the real time requirement for GPS/INS integrated systems during the firing stage of missiles.

关 键 词：组合导航 阵列结构 并行算法 GPS 

分 类 号：V249.328[航空宇航科学与技术—飞行器设计]